The P/E ratio
The most used valuation measure in the world, and the most misread. It is not a measure of cheapness. It is a measure of expectations.
The price-to-earnings ratio is a share price divided by earnings per share. If a share costs $100 and the company earned $5 per share last year, the P/E is 20.
One useful way to read it: at a P/E of 20, you are paying twenty dollars for each dollar of annual profit. If profits never changed and were all paid out to you, it would take twenty years to get your money back.
Three P/E ratios, one company
The denominator is contested, and the choice changes the answer dramatically.
NVIDIA at $225.16 per share (August 15, 2026):
÷ trailing FY2026 GAAP EPS of $4.90 = 45.9×
÷ annualised Q1 FY2027 non-GAAP EPS of $7.48 = 30.1×
÷ annualised Q1 FY2027 GAAP EPS of $9.56 = 23.6×
Same price. Same company. Same day. A factor-of-two difference depending on which earnings figure is used. This is why 'the P/E' is never a sufficient description.
Trailing versus forward
- Trailing P/E uses earnings already reported. It is factual but backward-looking, and for a fast-growing company it will look expensive even when it is not.
- Forward P/E uses estimated future earnings. It is more relevant and less reliable, because the estimate can be wrong. The S&P 500's forward 12-month P/E was 20.0 in early August 2026, against a ten-year average of 19.0.
Why a low P/E is not 'cheap'
This is the misunderstanding that costs beginners the most money.
A P/E ratio compares price to current earnings. If the market expects those earnings to fall, it will pay less for them — producing a low P/E. Cyclical businesses at the peak of their cycle routinely show their lowest P/E ratios at the moment they are most expensive in reality, because the earnings in the denominator are about to collapse.
A P/E ratio is a statement about expected growth and expected durability. A high number means the market expects earnings to grow, or to persist for a long time, or both. A low number means the opposite.
The question is never 'is this P/E high or low.' It is 'do I agree with the expectation embedded in it.'
When P/E does not work at all
- Companies with no earnings. The ratio is undefined.
- Companies with one-off gains or charges distorting the denominator.
- Banks and insurers, where price-to-book is often more informative.
- Any comparison across industries — a utility and a software company have structurally different multiples for structural reasons.